Depth Fusion Imaging for Autonomous Cleaner Distance Accuracy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Autonomous cleaners with 3D depth sensors face high costs and computational complexity, leading to low frame rates and resolution of images with absolute depth information, causing inaccurate distance determination during fast movement.

Innovation Solution

An image generation device and method that utilize multiple image sensors and processing circuits to generate images with first and second resolutions, employing depth estimation and fusion engines to create output images with first absolute depth information, reducing computational effort and cost.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a 3D depth sensor is used to generate images with absolute depth information, then distance determination accuracy is improved, but device cost and computation complexity increase

Engineering Contradiction:

Improvedistance determination accuracyVSAvoidcomputation complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the image processing task into multiple parts: using a 3D depth sensor for a portion of the image (second image with absolute depth information) and using a depth estimation engine for another portion (third image with relative depth information). This segmentation allows the system to achieve accurate distance determination in key areas while reducing overall computation complexity by processing only essential regions with high computational cost.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces a fusion engine as an intermediary component that combines the second image (with absolute depth information from the 3D depth sensor) and the third image (with relative depth information from the depth estimation engine) to generate the final output image. This intermediary approach allows the system to leverage the cost-effective depth estimation engine while maintaining the accuracy benefits of the 3D depth sensor in critical areas.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If a 3D depth sensor is used to generate images with absolute depth information, then distance determination accuracy is improved, but frame rate and resolution decrease

Engineering Contradiction:

Improvedistance determination accuracyVSAvoidframe rate

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The patent applies partial action by using the computationally intensive 3D depth sensor only for generating the second image with absolute depth information, rather than processing the entire image stream through it. The depth estimation engine processes additional images to generate the third image with relative depth information. This partial application of the 3D depth sensor maintains frame rate and resolution while ensuring accurate distance determination in critical areas.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The patent segments the image processing workload by dividing images into different categories: some images are processed through the 3D depth sensor (second images) and others through the depth estimation engine (first images). This segmentation allows the system to maintain high frame rates by processing images through the faster, less computationally intensive depth estimation engine while still achieving accurate distance determination where needed.

Inventive Principle:

Principle #1Segmentation

3Measurement precision

If a 3D depth sensor is used to generate images with absolute depth information, then distance determination accuracy is improved, but computation effort increases

Engineering Contradiction:

Improvedistance determination accuracyVSAvoidcomputation effort

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the computation effort by dividing image processing into two paths: one using the 3D depth sensor for absolute depth information (second images) and another using the depth estimation engine for relative depth information (first images). This segmentation reduces overall computation effort by avoiding the need to process every image through the more computationally intensive 3D depth sensor, while still maintaining accurate distance determination in critical areas.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The fusion engine acts as an intermediary that efficiently combines the second image (with absolute depth information) and the third image (with relative depth information) to generate the output image. This intermediary approach reduces computation effort by leveraging the pre-processed depth information from both sources rather than reprocessing the entire image data through the same computational pipeline.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12620063B2Image generation device, autonomous cleaner and image generation method

Publication Date: 2026.05.05 PIXART IMAGING INC

AI summary

An image generation device, for generating an output image with first absolute depth information, comprising: at least one image sensor, configured to generate at least one first image with a first resolution and at least one second image with a second resolution, wherein the first image does not have depth information and the second image has second absolute depth information, wherein the first resolution is higher than the second resolution; and a processing circuit, configured to use a depth estimation engine to generate at least one third image with a third resolution according to the first image, wherein the third image has relative depth information; wherein the processing circuit further uses a fusion engine to fuse the second image and the third image to generate the output image.